Senior Counsel
August 8, 2025
For the past decade, lawyers adopting best practices in technology-assisted review (TAR) have been guided by one overriding concern: defensibility. Their workflows were built to withstand judicial scrutiny as prescribed by the rules, case law and the Sedona Canada Principles.
That risk-averse approach serves its purpose. But it also entrenches a mindset that treated technology as a liability to be justified rather than as a capability to be unleashed. Today, generative AI invites us once again to shift our professional perspective from defensive compliance to proactive ethical design.
Beyond Defensibility: Toward Trustworthy AI Systems
In the world of generative AI, “defensibility” will no longer need to be treated as a constraint. Instead, it can become a strength, something built directly into the design of the tools we use. The increasing reliability of AI tools will open the door to a more ethical, transparent, and trustworthy form of eDiscovery. Here’s how:
Provenance Tracking: Every AI-generated summary, insight, or recommendation could be directly tied back to the original documents from which it was derived. This ensures that users, including the courts, can always verify where the information came from. As a result, unreliable or fabricated content (“hallucinations”) becomes easier to detect and eliminate, leaving only traceable, evidence-based outputs.
Explainability-by-Design: Instead of simply labelling a document “relevant” or “privileged,” AI tools could explain why those conclusions were reached. The reasoning is made clear, and linked to the supporting material. This level of transparency would allow legal professionals to better evaluate and trust the technology, and to make more informed decisions.
Human Oversight: Generative AI should interact with the judgment of lawyers. Every suggestion or answer generated by the system would remain subject to human review, challenge, and refinement. (And, I would argue, vice-versa.) This ensures that ultimate responsibility and accountability stay where they belong: with legal counsel.
With these capabilities, we can move beyond the traditional mindset of defensively justifying what was missed in review. Instead, the focus is on how we can build a clear, well-supported understanding of the facts. That’s a profound shift, and one that positions generative AI not just as a tool for efficiency, but as a foundation for better, fairer discovery.
What This Disruption Achieves
Effectively implemented, generative AI has the potential to radically reimagine what discovery is and how it serves the interests of the justice system. Here’s what that could look like:
1. Speed: What once took weeks of linear document review can be accomplished in minutes.
2. Insight: Discovery becomes strategic and investigative, rather than mechanical.
3. Accessibility: Smaller firms and under-resourced parties gain the analytic capacity to challenge large-scale document productions.
4. Cost-Efficiency: Volume-driven billing is reduced, and instead is focused on value-driven results.
5. Fairness: When technology is designed to clarify rather than check the boxes, we reduce the opportunity for tactical gamesmanship and return focus to the merits of the case.
The Barriers Ahead
Of course, the road to transformation can’t be easy. The legal profession remains inherently conservative: rooted in precedent, wary of risk, and typically slow to adopt innovative solutions. Many lawyers are rightly cautious, concerned about the reliability, admissibility, and ethical implications of AI-generated outputs in high-stakes litigation. Courts and regulators, too, are just beginning to consider how these technologies should be governed.
Moreover, entrenched business models within the legal ecosystem can reinforce the status quo. Traditional document review has become a predictable revenue stream for firms and service providers alike. A move toward AI-driven insight, which compresses timelines and reduces billable volume, can challenge financial incentives built around linear workflows and headcount-based review. For real change to take root, law firms, clients, and vendors alike must be willing to rethink not only their tools, but their business models and their professional culture.
A New Ethos for Legal AI
As a firm whose practice is dedicated exclusively to electronic discovery, we believe the future lies not in adapting generative AI to fit old workflows, but in rethinking the very architecture of eDiscovery.
This is more than a technical upgrade. It’s a call to rebuild discovery around trust, transparency, and human oversight. Done right, we don’t just make the process faster or cheaper – we make it better.
